Related posts:

1. Buying in London
2. World’s Best Bars 2011

1. Start with a base word. This can be anything, but the weirder the better. This will stay consistent across your passwords. For the purposes of explaining this technique, the base word will be "Redux".

2. Next take the name of the website or service. Let's say we are creating a password for Reddit. Take a number of specific letters from the name. Let's say that we always take the first 2 letters from the site name. This gives us the letters 'R' and 'e'. If you want, instead of taking the first 2 letters you could take the first and last letter, or the first 3 letters. It doesn't matter. What matters is to always take the same number of characters from the same position of each website.

3. Next, use an algorithm. This algorithm will be unique to you, but is a way of mixing the letters you have extracted from the website into your base word. A simple algorithm might be to prepend the letters to your base word. Your password then becomes "ReRedux". A slightly better algorithm would be to increment the letters by one character. This means that "Re" -> "Sf". Then merge the resultant characters into the first and third positions of your base word. Your password for Reddit then becomes "SRfedux".

This combination of using a unique base word, that is consistent across all your passwords, but modifying it using an algorithm based on the website name will allow you to construct passwords that you can remember, but are different for each site you need to log in to.

EDIT: OK, based on the Reddit comment thread it turns out that I didn't properly explain the rationale for using a combination of the website name and some constant stream of characters is to try and create a unique password for each individual site. You want to have a unique password per site because a lot of sites store your password in plain-text. If one site gets cracked, you don't want the attackers to be able to compromise your identity on other sites you have an account on. Also, using a passphrase or the first letters of words in a passphrase as your base word is obviously more secure than using a dictionary word.

EDIT 2: It seems people are worried about the entropy of the generated passwords. Obviously, they are not completely random, but to give you an idea on how secure they can appear, here are some passwords generated from the same base word and algorithm for different sites:

The algorithm that generated these passwords was fairly simple. A slightly more complex algorithm would generate different length passwords. The key takeaway is that if an attacker gets hold of the password for the reddit site, it is extremely unlikely that they will be able to reverse-engineer the passwords for the other sites.

Related posts:

1. Written by Robot

Throughout 2011 though, I have written a paragraph each and every day in my iPhone. Kind of a mini-diary. It's been the longest period of regular writing that I have ever had. The factor that's made me write regularly for so long is that I put a repeating reminder in my iPhone. At 4:45pm each day my phone rings a bell, which reminds me that I need to write down my Momento for the day.

Now I'm trying to apply this same technique to my blogging - setting aside a time to write, and writing a blog post each day.

Capturing Information

I use Evernote to capture information throughout the day, especially interesting articles I read online. Within Safari, I can highlight the text I want to capture, press Cmd-C to copy it, and then press Ctrl-Cmd-V to paste it to a new note in Evernote. It's fast, and it also saves the URL for attribution

Emacs - org-mode and markdown

When I'm ready to write, I start off by creating a directory to write in. In fact, I have an Emacs function that creates the directory for me, along with an org-mode file, and a markdown file. The org-mode file is where I write the skeleton of the post, and the markdown file is where I write the prose. Here is the emacs lisp code that sets me up for blogging:

 
(defvar bh-blog-article-dir (expand-file-name "~/Dropbox/docs/blog-posts"))
 
(defun bh-blog-article-start ()
  (interactive)
  (let ((topic-name (read-string "Topic: "))
        (post-dir nil)
        )
    (setq base-post-topic (mapconcat (function (lambda (x) x))  (split-string topic-name) "-"))
    (setq post-dir  (concat bh-blog-article-dir "/" (format-time-string "%Y%m%d") "-" base-post-topic))
    (make-directory post-dir)
    (find-file (concat post-dir "/" base-post-topic ".org"))
    (find-file-other-window (concat post-dir "/" base-post-topic ".md"))
    )
  )
 

Uploading

When a post is ready to be uploaded, I have a Python script that converts the markdown file to HTML format, extracts the first H1 heading in the HTML to use as the post title, and then uploads the post to my blog.

Related posts:

1. Integrating Emacs and XCode
2. Opening a header file in Emacs
3. Emacs and Kanban

No related posts.

I want to start treating my blog as though I was a gardner tending his garden. Regularly adding new posts. Replying quickly to any comments. Upgrading and improving my blogging environment as I can. I would like to minister to a site that will encourage people to stroll around and smell the flowers.

Related posts:

1. Blogging Process
2. Blogging and SEO talk by Matt Cutts

Related posts:

1. Blogging as Gardening
2. Blogging Process
3. Dr Jill Bolte Taylor’s stroke experience

The core model for viral growth is the following: viral coefficient = (average number of users invited by each active user) x (proportion of invited users that actually join or become active) x (proportion of such users that invite others). Daniel provides a checklist of techniques designed to optimize the viral coefficient. The general techniques are as follows:

• Make inviting people a core part of the process.
• Keep pulling people back in
• Be useful even if there are no other friends using the application.
• Remove artificial invitation limits.

Related posts:

1. Cool hand Luke